CN112565082B - Service chain mapping method based on hybrid network, intelligent terminal and storage medium - Google Patents

Abstract

The invention discloses a service chain mapping method based on a hybrid network, an intelligent terminal and a storage medium, wherein the method comprises the following steps: acquiring a service link request; determining a target candidate path set corresponding to the service link request according to a preset static path table construction rule and preset network link information; calculating a path evaluation value corresponding to each target candidate path according to a preset network real-time index; and determining the target transmission path in the target candidate path set according to the path evaluation value. The invention dynamically selects the target transmission path corresponding to the service link request based on the network real-time index of each node in the current network, so that the service link can reasonably and effectively schedule and allocate limited network resources.

Description

Service chain mapping method, intelligent terminal and storage medium based on hybrid network

technical field

The present invention relates to the field of computer technology, in particular to a hybrid network-based service chain mapping method, an intelligent terminal and a storage medium.

Background technique

Traditional networks often require a large number of network functions (Network Fuction, NF) deployed on specific devices, including load balancers, firewalls, network address translation, etc., to provide special services. These devices are called network middleware. In the network function virtualization (Network Function Virtualization) technology, a network function can be virtualized into a virtual network function (Virtual Network Function, VNF), and each VNF is deployed on each physical node in the form of software. Traffic is processed by a series of different VNFs in the order of demand, forming a service chain. The link between two VNFs is called a virtual link, and the link between two physical nodes is called a physical link. After each VNF is deployed on each physical node, in order to ensure the normal operation of the service chain, it is also necessary to map the virtual links between the various VNFs in the service chain to the physical paths between the physical nodes, that is, from a physical node A path formed by a link to another physical node.

In traditional networks, physical nodes where VNFs are deployed are usually general-purpose servers. The network function node includes a server and a programmable data plane. The server implements network functions in the form of software VNFs, while the programmable data plane is a hardware network function based on a match-action table. In recent years, many studies have proposed to offload network functions using the P4 programmable data plane. The P4 programmable data plane has higher processing power than general-purpose servers and can provide higher performance for traffic, and by configuring the match-action table in the programmable data plane pipeline, it can provide programmability for offloading Network features. For example, the network address translator can match the traffic IP address through the data plane match-action table and modify it to implement the IP address mapping function for the given IP address. Multiple network functions can usually be installed on a P4 programmable data plane to provide network function services for corresponding flows by issuing flow tables. The P4 programmable data plane has its own limitations and cannot support some network functions involving complex operations (including operations such as reading load data). The hardware resources also need to ensure the performance of traffic forwarding, so programmable switches and The environment in which the general server is combined provides network function services.

Traditional service chain mapping algorithms usually only consider pure server VNF scenarios. However, because the P4 programmable data plane NF has the following characteristics, the traditional model is difficult to directly apply to the hybrid system:

(1) Unlike the server VNF, the traffic passes through the programmable data plane pipeline sequentially, so it must also pass through the NFs on the programmable data plane in a certain order. If the two NFs on the same switch are passed in the reverse order, additional ports will be occupied and the performance of other traffic will be greatly affected;

(2) Different from the server VNF, when the traffic passes through the programmable switch NF, it mainly affects the use of memory resources, so the algorithm needs to consider the programmable switch resource limit when making decisions.

At the same time, due to the need to provide real-time decision-making for dynamic traffic in the network, the solution should make timely decisions and provide quick responses, thereby reducing the impact on traffic delays. For the service chain mapping scheme, simply using the shortest path value between two nodes for VNF mapping will lack consideration of the real-time status of the network and easily cause congestion; there are schemes by constructing a service forwarding graph and running the shortest path value in real time. The algorithm adapts to the dynamic nature of the network, but the time complexity of the shortest path value algorithm is high, which increases the network traffic delay. Therefore, the service mapping algorithm needs to cope with the changing traffic demand, quickly select the appropriate network function instance for the traffic and construct the service forwarding path, so as to reduce the flow completion time and avoid the generation of network congestion.

SUMMARY OF THE INVENTION

The main purpose of the present invention is to provide a hybrid network-based service chain mapping method, an intelligent terminal and a storage medium, aiming to solve the problem of slow response speed of hybrid network-based service chain mapping in the prior art.

In order to achieve the above object, the present invention provides a hybrid network-based service chain mapping method, the hybrid network-based service chain mapping method includes the following steps:

Get service link request;

Determine the target candidate path set corresponding to the service link request according to the preset static path table construction rule and preset network link information;

Calculate the path evaluation value corresponding to each of the target candidate paths according to preset network real-time indicators;

According to the path evaluation value, a target transmission path in the target candidate path set is determined.

Optionally, the hybrid network-based service chain mapping method, wherein the target candidate path corresponding to the service link request is determined according to a preset static path table construction rule and preset network link information. set, including:

constructing a network forwarding graph corresponding to the service link request according to a preset forwarding graph rule and the network link information;

determining a target candidate path set corresponding to the service link request according to the network forwarding graph;

Wherein, the service link request is R=(N _r ,v _s ,v _d ), N _r is a sequence composed of network functions corresponding to the service link request, and v _s is the service link request that issued the source node, where v _d is the target node corresponding to the service link request;

The network forwarding graph G=(N, V, E), where V is the set of network nodes, N is the network function corresponding to each of the network nodes, and E is obtained by connecting the network nodes according to the network link. The network forwarding graph includes |N _r |+2 network stages, |N _r | is the length of N _r , where the first stage is the source node, and the |N _r |+2 stage is the target node;

V_i为网络功能转发在第i网络阶段的网络节点的集合，

为第i网络阶段的各个网络节点具备的网络功能，

为具备网络功能

的网络节点的集合。the network node

V _i is the set of network nodes forwarded by the network function in the i-th network stage,

is the network function possessed by each network node in the i-th network stage,

for network function

A collection of network nodes.

Optionally, in the hybrid network-based service chain mapping method, the edge includes an edge weight, and the edge weight includes a shortest path value between the network nodes.

Optionally, in the hybrid network-based service chain mapping method, the edge is a directed edge, and the start node of the edge points to a termination node, and the start node is a network corresponding to the termination node The previous stage of the stage; the calculation process of the edge weight specifically includes:

For each of the edges, when the start node and the end node of the edge are located on different switches, assign the shortest path value corresponding to the start node and the end node to the edge weight corresponding to the edge;

When the start node and the end node corresponding to the edge are located in the same physical node, determine whether the network function installation sequence in the physical node is from the start node to the end node;

If so, assign zero to the edge weight corresponding to the edge;

If not, assign the preset invalid weight value to the edge weight corresponding to the edge.

Optionally, in the hybrid network-based service chain mapping method, the determining a target candidate path set corresponding to the service link request according to the network forwarding graph specifically includes:

determining an initial candidate path set corresponding to the service link request according to a preset shortest path value algorithm and the network forwarding graph;

According to a preset screening rule, the initial candidate path set is screened to obtain a target candidate path set.

Optionally, in the hybrid network-based service chain mapping method, wherein the initial candidate path set is screened according to a preset screening rule to obtain a target candidate path set, which specifically includes:

According to the order from small to large and the preset number of target candidate paths, the initial candidate paths that do not include loops in the initial candidate path set are sequentially selected as target candidate paths;

According to the target candidate path, a corresponding target candidate path set is generated.

Optionally, in the hybrid network-based service chain mapping method, wherein, the service link request further includes traffic rate and required resources, the network real-time indicator includes remaining bandwidth and remaining resources, and the The path evaluation value is the sum of the link evaluation values corresponding to each edge in the target candidate path.

Optionally, in the hybrid network-based service chain mapping method, the calculating a path evaluation value corresponding to each target candidate path according to a preset network real-time index specifically includes:

For each edge in the target candidate path, determine whether the edge satisfies a preset resource constraint;

If so, calculate the ratio of the remaining bandwidth corresponding to the edge to the traffic rate, and use the ratio as the link evaluation value corresponding to the edge;

If not, the preset invalid link value is used as the link evaluation value corresponding to the edge.

Optionally, in the hybrid network-based service chain mapping method, the resource constraint condition is whether the remaining bandwidth of the network link corresponding to the edge is less than the traffic rate, and the starting node corresponding to the edge has a Whether the remaining resources are greater than the required resources.

In addition, in order to achieve the above object, the present invention also provides an intelligent terminal, wherein the intelligent terminal includes: a memory, a processor, and a hybrid network-based service stored in the memory and running on the processor A chain mapping program, when the hybrid network-based service chain mapping program is executed by the processor, implements the steps of the hybrid network-based service chain mapping method as described above.

In addition, in order to achieve the above object, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a hybrid network-based service chain mapping program, and the hybrid network-based service chain mapping program is The processor implements the steps of the hybrid network-based service chain mapping method described above when executed.

The invention proposes a service chain mapping method based on a hybrid network. First, according to the service link request, the corresponding target candidate path set is determined, and then the network real-time index is used as the evaluation standard to screen, and the optimal path in the target candidate path set is selected as the target candidate path set. target transmission path, thereby optimizing network performance and avoiding congestion as much as possible. The present invention is oriented towards a hybrid network function including a P4 programmable data plane network function and a software network function, provides fast and flexible automatic service path construction for network traffic, while shortening response time, reducing flow completion delay and improving network performance.

Description of drawings

1 is a flowchart of a preferred embodiment provided by the hybrid network-based service chain mapping method of the present invention;

2 is a flow chart between hardware in a service chain mapping system in a preferred embodiment provided by the hybrid network-based service chain mapping method of the present invention;

3 is a schematic diagram of a service chain mapping system architecture in a preferred embodiment provided by the hybrid network-based service chain mapping method of the present invention;

4 is a schematic diagram of a network forwarding diagram in a preferred embodiment provided by the hybrid network-based service chain mapping method of the present invention;

5 is a schematic diagram of judging the first node and the second node in the preferred embodiment provided by the hybrid network-based service chain mapping method of the present invention;

6 is a static path construction algorithm in a preferred embodiment provided by the hybrid network-based service chain mapping method of the present invention;

7 is a dynamic path selection algorithm in a preferred embodiment provided by the hybrid network-based service chain mapping method of the present invention;

FIG. 8 is a schematic diagram of an operating environment of a preferred embodiment of an intelligent terminal of the present invention.

Detailed ways

In order to make the objectives, technical solutions and advantages of the present invention clearer and clearer, the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention, but not to limit the present invention.

The hybrid network-based service chain mapping method according to the preferred embodiment of the present invention, as shown in FIG. 1 , the hybrid network-based service chain mapping method includes the following steps:

Step S100, acquiring a service link request.

Specifically, the hybrid network-based service chain mapping method in this embodiment is applied to the service chain mapping system. As shown in Figures 2 and 3, the service chain mapping system includes a data plane and a control plane. The data plane includes software virtualized network function instances installed on servers and hardware network function instances installed on programmable switches. The control plane consists of It consists of decision-making layer, command translation module and information collection module including service chain mapping algorithm. The hardware network function instance of the data plane is implemented based on a match-action table (Match-Action Table) element, and is activated by corresponding rules issued by the control plane at runtime. Traffic can selectively pass through one or more network functions as it passes through a switch, as indicated by the flow table rules. Software virtualized network function instances are implemented on servers to carry network functions that are too complex to be supported by the primitives provided by the programmable data plane.

When the data plane recognizes the incoming new data flow, its first data packet will be transmitted to the control plane as a service link request for service path construction, that is, the service chain request in the service chain request collection in Figure 2 . The information collection module of the control platform provides an interface for data plane communication, and through the information collection module, the control platform obtains the service link request. In addition, the information collection module will also pre-determine the link information required by the decision-making layer and the network test real-time indicators. The link information includes each network node in the network, the link relationship between the network nodes, and the network services provided by each network node, etc. , the network test real-time indicators include the network bandwidth occupancy rate, the processing speed of each network node, and other nodes used to evaluate the processing capability of the network nodes.

Step S200, according to a preset static path table construction rule and preset network link information, determine a target candidate path set corresponding to the service link request;

Specifically, since the information collection module also determines the link information required by the layer in advance, it can determine the possible link information according to the network link information, the source node, the target node and the required network service in the service link request. target candidate paths, and use all target candidate paths as the target candidate path set. A rule for constructing a static path table is preset in order to determine the target candidate path set corresponding to the link request according to the current static network environment, that is, network link information. In the process of determining its possible target candidate paths, in order to reduce the amount of calculation, certain screening can be used, for example, the top k shortest path value algorithm is used to select several paths with the shortest paths as target candidate paths.

Further, referring to FIG. 4 , in the process of determining the target candidate path set corresponding to the service link request in this embodiment, a static path construction algorithm is adopted, and the specific process is as follows:

A10. Construct a network forwarding map corresponding to the service link request according to a preset forwarding map rule and the network link information.

Specifically, after obtaining the service link request, the decision layer of the control platform constructs a network forwarding graph corresponding to the service link request according to a preset forwarding graph rule, wherein the forwarding graph rule is a rule for constructing a network forwarding graph, The network forwarding graph is constructed according to the link information and network requests collected before by the information collection module.

The forwarding graph rule is used to filter out the paths in the current network that can satisfy the basic network function requested by the service link. The starting point of the network forwarding graph, that is, the source node, is the node that sends the service link request, the end point, that is, the target node, is the target node corresponding to the service link request, and the middle node of the network forwarding graph is A node based on the network service requirements listed in the service link request. Therefore, any path in the network forwarding graph can satisfy the most basic service requirement of the service link request.

The service link request is defined as consisting of a source node, a target node and a network function sequence, that is, R=(N _r ,v _s ,v _d ), and N _r is the network function corresponding to the service link request The network function is the service requirement corresponding to the service link request, v _s is the source node that sends the service link request, and v _d is the target node corresponding to the service link request.

的交换机或服务器节点组成，表示为

即：

V_i为网络功能转发在第i阶段的网络节点的集合，

为第i阶段的各个网络节点具备的网络功能，

为具备网络功能

的网络节点的集合。The network forwarding graph G=(N, V, E), where V is the set of network nodes (including servers and switches), N is the network function corresponding to each of the network nodes, and E is the network link according to the network link. The edges obtained by connecting the network nodes, the network forwarding graph includes |N _r |+2 segments, |N _r | is the length of N _r , where the first stage and the last stage are the source node and target node of the flow, and the remaining The node v in each phase i is installed by the corresponding network function

composed of switches or server nodes, expressed as

which is:

V _i is the set of network nodes forwarded by the network function in the i-th stage,

is the network function possessed by each network node in the i-th stage,

for network function

A collection of network nodes.

Further, the edge includes an edge weight, and the edge weight includes a shortest path value between the network nodes.

Specifically, as shown in FIG. 4 , the weight g[v _{i -1} ][vi ] of the edge in the forwarding graph is set as the shortest path value d[v between the network node v and the network node u in the network function forwarding graph _i-1 ][v _i ], the shortest path value d[v _i-1 ][v _i ] can be calculated based on network link information, and is usually obtained from an application programming interface (Application Programming Interface, API) provided by the control plane. Generally, the shortest path value refers to the number of hops between network nodes in the topological sense. If they are directly connected, the shortest path value is 1; if they are connected through a node, the shortest path value is 2.

Further, in order to reduce the end-to-end delay of the flow, in this embodiment, the shortest path value is adjusted on the number of hops between network nodes, and the specific process is as follows:

For each of the edges, when the start node and the end node of the edge are located on different switches, assign the shortest path value corresponding to the start node and the end node to the edge weight corresponding to the edge;

When the start node and the end node corresponding to the edge are located in the same physical node, determine whether the network function installation sequence in the physical node is from the start node to the end node;

If so, assign zero to the edge weight corresponding to the edge;

If not, assign the preset invalid weight value to the edge weight corresponding to the edge.

Specifically, the edge is a directed edge, and the start node of the edge points to the end node, vi _-1 is the start node, _vi is the end node, and the start node is the network corresponding to the end node the previous stage of the stage. For each of the edges, when the start node and the end node corresponding to the edge are located in different switches, the edge weight corresponding to the edge is still the shortest path value between the two network nodes.

Referring to Figure 5, if v _i-1 and v _i are the same physical node in the network, such as a switch node, the weights should be set according to the following two cases: (1) If the switches corresponding to v _i-1 and v _i The network functions of the corresponding stages of v _i-1 and v _i are installed in a sequential manner, the data packets can pass through the two network functions in a pipelined manner, and there is no path cost. Therefore, the edge weight is assigned to 0; (2 ) If the switches corresponding to v _i-1 and v _i are installed with the network functions of the corresponding stages of v _i-1 and v _i in reverse order (that is, the control logic on the switch is to first pass the network functions of the corresponding stage of u and then pass through the corresponding stages of v ) stage network function), that is, the weight value of the edge is set to a particularly large value, that is, an invalid weight value, to prevent this path from being selected. The reason for this setting is that although it is possible to pass the network function in reverse order by resubmitting the data packets to the starting point of the pipeline on the same switch, there is no path cost, but because the resubmitted traffic additionally occupies the ingress port, it will be It has a great impact on the performance of other traffic on the switch and is not conducive to the overall network performance. The formula for this process can be expressed as:

则表示v_i对应物理节点上，是否以从

到

的顺序安装网络功能(对于安装有对应网络功能的服务器节点，此变量值恒为1)，MAXIMUM即无效权重值。Among them, p(n) represents the physical node corresponding to the network node u,

It means that on the physical node corresponding to v _i , whether the

arrive

The network functions are installed in the order of (for the server node installed with the corresponding network function, the value of this variable is always 1), and MAXIMUM is the invalid weight value.

A20. Determine a target candidate path set corresponding to the service link request according to the network forwarding map.

Specifically, since the network forwarding graph includes the most basic service requirements in the current network that can meet the service link request, each path in the network forwarding graph is used as a target candidate path, so as to determine the corresponding service link request. The set of target candidate paths.

Further, based on the above-mentioned network forwarding graph, this embodiment collects and determines the target candidate path set in a manner based on K shortest paths (k-shortest paths, KSP) algorithm.

B10: Determine an initial candidate path set corresponding to the service link request according to a preset shortest path value algorithm and the network forwarding graph.

Specifically, the KSP algorithm is first used to determine the initial candidate path set corresponding to the service link request based on the network forwarding graph. The KSP algorithm adopts the idea of the deviation path algorithm in the recursion method, and is suitable for the directed acyclic graph structure with non-negative weight edges. The specific process is:

According to the shortest path value algorithm and the edge weights corresponding to each of the edges, an initial candidate path corresponding to the network forwarding graph and a path weight value corresponding to each initial path are generated;

According to the path weight value, sort the initial candidate paths from small to large to generate a sequence of candidate paths;

Selecting the top m initial candidate paths in the sequence of candidate paths as the initial candidate path set, where m is a preset number of initial candidate paths.

Specifically, the first shortest path value P(1) is calculated first, and then the other K-1 shortest path values are calculated sequentially on this basis. When calculating P(i+1), all nodes on P(i) except the target node are regarded as deviating nodes, and the shortest path value from each deviating node to the terminal node is calculated, that is, the above-mentioned edge weight, and then It is spliced with the path from the starting node to the deviating node on the previous P(i) to form an initial candidate path and calculate its corresponding path weight value to generate a candidate path sequence. Then, according to the path weight value, each initial candidate path is sorted from small to large, and finally according to the preset number m of initial candidate paths, the first m initial candidate paths in the sequence of candidate paths are selected as the initial candidate path set. Referring to FIG. 6 , where m=2k, k is the preset number of target candidate paths. Except for 2k, m can be a positive integer greater than k, as long as the subsequent screening is satisfied, there is an amount of data that satisfies the screening.

B20: Screen the initial candidate path set according to a preset screening rule to obtain a target candidate path set.

Specifically, a screening rule is preset, and the screening rule is used to screen the initial candidate path set. The screening rule can be based on whether the initial candidate path contains loops and the number of loops, for example, a loop quantity threshold is preset. , if the loop quantity of the initial candidate path exceeds the loop quantity threshold, delete the initial candidate path.

Further, referring to Fig. 6, the screening rule is for whether the initial candidate path contains a loop, and the specific process is as follows:

According to the order from small to large and the preset number of target candidate paths, the initial candidate paths that do not include loops in the initial candidate path set are sequentially selected as target candidate paths;

According to the target candidate path, a corresponding target candidate path set is generated.

Specifically, because the existence of loops will lead to the problem of route conflict on the path intersection node, and it is difficult to distinguish the next hop position for the conflicting traffic in the current system, it is necessary to perform path screening to avoid the existence of loops.

Then, according to the order from small to large, the initial candidate paths that do not contain loops are taken as target candidate paths and added to the target candidate path set. When the number of target candidate paths reaches k, the algorithm stops and the target candidate path set is obtained.

Step S300: Calculate path evaluation values corresponding to each of the target candidate paths according to preset network real-time indicators.

Specifically, in this embodiment, a dynamic path selection algorithm is used to select the target transmission path among the target candidate paths. The path evaluation value refers to a value calculated according to the network real-time indicators such as the network status and resource utilization of each target candidate path, which is used to evaluate the transmission capability of each target candidate path. Since the control platform obtains the network real-time indicators through the information gathering module, it can calculate and evaluate the transmission capacity of each target candidate path in the target candidate path set according to the network real-time indicators, and obtain the corresponding path evaluation value.

Further, in this embodiment, the link bandwidth utilization rate is used as the path evaluation value, the service link request further includes the traffic rate and the required resources, the network real-time indicator includes the remaining bandwidth and remaining resources, and the path evaluation value is The sum of the link evaluation values corresponding to each edge in the target candidate path.

其中，m_p为目标候选路径p对应的路径评价值，b_r表示请求R的发送速率，b_e表示路径p包含的一段物理链路e的剩余带宽，e为目标候选路径p对应的链路评价值。求和结果即为累计链路带宽利用率。The formula for the path evaluation value is

Among them, mp is the path evaluation value corresponding to the target candidate path _p , br represents the sending rate of the request _R , _bee represents the remaining bandwidth of a physical link e included in the path p, and e is the link corresponding to the target candidate path p Evaluation value. The summation result is the cumulative link bandwidth utilization.

Referring to Figure 7, the dynamic algorithm module obtains the link load information and node resource usage from the data plane in real time through the information collection module. Feasibility, the method for evaluating whether the node capacity and the link capacity are satisfied in this embodiment is to calculate the link evaluation value corresponding to the link. The preferred calculation process of the link evaluation value in this embodiment is:

Step C10, for each edge in the target candidate path, determine whether a preset resource constraint condition is satisfied.

且

其中，c_r表示服务链路请求R所需求的资源，c_e表示链路e起点所包含的剩余资源，也就是链路e对应的边的起始节点的剩余资源。Specifically, traverse the edges of each target candidate path to determine whether the edge satisfies the resource constraints, that is, whether the remaining bandwidth of the network link corresponding to the edge is less than the traffic rate, and whether the remaining resources of the starting node corresponding to the edge are greater than the required resources. The resource constraints adopted in this embodiment are: the target transmission path is required to satisfy

and

Among them, cr represents the resources required by the service link request _R , and ce represents the remaining resources included in the starting point of link _e , that is, the remaining resources of the starting node of the edge corresponding to link e.

In addition to the above resource constraints, the resource ratio threshold, the ratio of computing demand resources to remaining resources, and the relationship between the resource ratio threshold, etc., and the utilization of bandwidth and resources can also be used as resource constraints. According to the node type, the physical nodes corresponding to the network nodes are classified as switches or servers. If it is a switch, the remaining resources refer to the remaining resources of the switch; if it is a server, the remaining resources are the remaining resources of the server, such as CPU resources. These two conditions can be regarded as resource constraints, and each target candidate path can be traversed according to the resource constraints.

Step C20, if yes, calculate the ratio of the remaining bandwidth corresponding to the edge to the traffic rate, and use the ratio as the link evaluation value corresponding to the edge.

并将其作为该边对应的链路评价值。Specifically, if yes, calculate the ratio of the remaining bandwidth corresponding to the edge to the traffic rate, that is,

And take it as the link evaluation value corresponding to the edge.

Step C30, if not, take the preset invalid link value as the link evaluation value corresponding to the edge.

Specifically, an invalid link value is preset, and the invalid link value may be a maximum value or infinite. If not, the preset invalid link value is used as the link evaluation value corresponding to the edge to prevent the target candidate path where the link is located from being selected.

Therefore, in this embodiment, considering the resource limitations and link bandwidth limitations of switches and servers, on the premise that the traffic demand can be met, the path with the smallest cumulative link bandwidth utilization rate is selected, so as to ensure network load balance, reduce network load, and enable service Reasonable and effective scheduling and allocation of limited network resources, reducing the deployment overhead of operators, improving the reception rate of requests, and achieving higher deployment benefits.

Step S400: Determine a target transmission path in the target candidate path set according to the path evaluation value.

Specifically, after obtaining the path evaluation value corresponding to each target candidate path, if the transmission capability is stronger and the path evaluation value is higher, the target candidate path corresponding to the path evaluation value with the largest value is selected as the target transmission path; if the transmission capability is stronger , the smaller the path evaluation value is, the target candidate path corresponding to the path evaluation value with the smallest value is selected as the target transmission path. In this embodiment, the stronger the transmission capability, the smaller the path evaluation value. Therefore, the target candidate path corresponding to the path evaluation value with the smallest value is selected as the target transmission path.

The decision-making layer sends the target transmission path as a deployment command to the command translation module to generate corresponding network function rules and traffic routing rules, and install them on the switches along the route. At the same time, the data packet is sent back to the data plane, and the data packet matches the flow table rules for normal forwarding. The subsequent data packets of the flow will be forwarded in the data plane according to the installed network function path. When the flow ends, the network function flow table and forwarding flow table rules will be automatically removed by the switch according to the timeout rule. Therefore, the target transmission path corresponding to each service link request is obtained by screening according to the latest network conditions, so that the service chain can reasonably and effectively schedule and allocate limited network resources.

In this embodiment, the target transmission network is quickly determined through a two-stage algorithm. The two-stage algorithm consists of a static path generation algorithm and a dynamic path selection algorithm. The static algorithm is called when the network is initialized, and the shortest path value is used as the index to filter out all the The set of candidate paths for the request; when the control plane receives the first data packet of the flow, the dynamic algorithm is invoked, and the target transmission path is selected from the set of target candidate paths for the request in combination with the network real-time indicators obtained from the information collection module.

The solution of this embodiment can be deployed in various ways. For example, it is deployed in the data center network to provide service chain mapping services for ultra-large-scale traffic, allocate resources, construct service paths, improve network performance and traffic forwarding efficiency, reduce traffic end-to-end latency, and save energy by offloading to the programmable data plane. Part of the server resources are used to better provide cloud services. It can also be deployed in the access network of the network provider, making full use of the resources and performance of programmable switches to simplify network deployment, build different virtual networks for different tenants and services, and improve service quality at the same time.

Further, as shown in FIG. 8 , based on the above hybrid network-based service chain mapping method, the present invention also provides an intelligent terminal correspondingly, and the intelligent terminal includes a processor 10 , a memory 20 and a display 30 . FIG. 8 only shows some components of the smart terminal, but it should be understood that it is not required to implement all the shown components, and more or less components may be implemented instead.

In some embodiments, the memory 20 may be an internal storage unit of the smart terminal, such as a hard disk or a memory of the smart terminal. In other embodiments, the memory 20 may also be an external storage device of the smart terminal, for example, a plug-in hard disk equipped on the smart terminal, a smart memory card (Smart Media Card, SMC), a secure digital (Secure) Digital, SD) card, flash memory card (Flash Card), etc. Further, the memory 20 may also include both an internal storage unit of the smart terminal and an external storage device. The memory 20 is used to store application software and various types of data installed in the smart terminal, such as program codes for installing the smart terminal. The memory 20 can also be used to temporarily store data that has been output or is to be output. In one embodiment, a hybrid network-based service chain mapping program 40 is stored on the memory 20, and the hybrid network-based service chain mapping program 40 can be executed by the processor 10, so as to realize the hybrid network-based service chain in this application. mapping method.

In some embodiments, the processor 10 may be a central processing unit (Central Processing Unit, CPU), a microprocessor or other data processing chips, which are used to execute program codes or process data stored in the memory 20, such as Execute the hybrid network-based service chain mapping method and the like.

In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, and the like. The display 30 is used for displaying information on the smart terminal and for displaying a visual user interface. The components 10-30 of the intelligent terminal communicate with each other through the system bus.

In one embodiment, when the processor 10 executes the hybrid network-based service chain mapping program 40 in the memory 20, the above hybrid network-based service chain mapping method is implemented.

The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a hybrid network-based service chain mapping program, and the hybrid network-based service chain mapping program is executed by a processor to achieve the above The steps of the hybrid network-based service chain mapping method described above.

Of course, those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing relevant hardware (such as processors, controllers, etc.) through a computer program, and the programs can be stored in a In the computer-readable storage medium, when the program is executed, the process of each method embodiment described above may be included. The computer-readable storage medium may be a memory, a magnetic disk, an optical disk, or the like.

It should be understood that the application of the present invention is not limited to the above examples. For those of ordinary skill in the art, improvements or transformations can be made according to the above descriptions, and all these improvements and transformations should belong to the protection scope of the appended claims of the present invention.

Claims (8)

1. A hybrid network-based service chain mapping method is characterized in that the hybrid network-based service chain mapping method comprises the following steps:

acquiring a service link request;

determining a target candidate path set corresponding to the service link request according to a preset static path table construction rule and preset network link information;

calculating a path evaluation value corresponding to each target candidate path according to a preset network real-time index;

determining target transmission paths in the target candidate path set according to the path evaluation values;

the determining a target candidate path set corresponding to the service link request according to the preset static path table construction rule and the preset network link information specifically includes:

constructing a network forwarding graph corresponding to the service link request according to a preset forwarding graph rule and the network link information;

determining a target candidate path set corresponding to the service link request according to the network forwarding graph;

wherein the service link request is R ═ (N)_r,v_s,v_d)，N_rCorresponding for said service link requestSequence of network function components, v_sV source node for issuing said service link request_dRequesting a corresponding target node for the service link;

the network forwarding graph G ═ N, V, E, (V is a set of network nodes, N is a network function corresponding to each network node, E is an edge obtained by connecting the network nodes according to a network link, and the network forwarding graph includes | N_rL +2 network stages, | N_rL is N_rWherein the first stage is the source node, the | N th stage_rThe stage of | plus 2 is a target node;

the network node

V_iThe set of network nodes at the ith network stage is forwarded for the network function,

for the network functions that are provided by the individual network nodes of the ith network stage,

to have network function

A set of network nodes of (a);

the edge comprises an edge weight comprising a shortest path value between the network nodes;

the edge is a directed edge, the starting node of the edge points to the terminating node, and the starting node is the previous stage of the network stage corresponding to the terminating node; the calculation process of the edge weight specifically includes:

for each edge, when the starting node and the terminating node of the edge are positioned in different switches, assigning the shortest path value corresponding to the starting node and the terminating node to the edge weight corresponding to the edge;

when the starting node and the ending node corresponding to the edge are positioned in the same physical node, judging whether the network function installation sequence in the physical node is from the starting node to the ending node;

if yes, assigning zero to the edge weight corresponding to the edge;

and if not, assigning a preset invalid weight value to the edge weight corresponding to the edge.

2. The hybrid network-based service chain mapping method according to claim 1, wherein the determining the target candidate path set corresponding to the service link request according to the network forwarding graph specifically includes:

determining an initial candidate path set corresponding to the service link request according to a preset shortest path value algorithm and the network forwarding graph;

and screening the initial candidate path set according to a preset screening rule to obtain a target candidate path set.

3. The hybrid network-based service chain mapping method according to claim 2, wherein the screening the initial candidate path set according to a preset screening rule to obtain a target candidate path set specifically comprises:

sequentially selecting initial candidate paths not including loops in the initial candidate path set as target candidate paths according to the sequence from small to large and the preset number of the target candidate paths;

and generating a corresponding target candidate path set according to the target candidate path.

4. The hybrid network-based service chain mapping method according to any one of claims 1 to 3, wherein the service link request further includes a traffic rate and a demand resource, the network real-time indicator includes a remaining bandwidth and a remaining resource, and the path evaluation value is a sum of link evaluation values corresponding to respective edges in the target candidate path.

5. The hybrid network-based service chain mapping method according to claim 4, wherein the calculating a path evaluation value corresponding to each of the target candidate paths according to a preset network real-time index specifically includes:

aiming at each edge in the target candidate path, judging whether the edge meets a preset resource constraint condition;

if so, calculating the ratio of the residual broadband corresponding to the edge to the flow rate, and taking the ratio as the link evaluation value corresponding to the edge;

and if not, taking the preset invalid link value as the link evaluation value corresponding to the edge.

6. The hybrid network-based service chain mapping method according to claim 5, wherein the resource constraint condition is whether the remaining bandwidth of the network link corresponding to the edge is smaller than the traffic rate and whether the remaining resource of the start node corresponding to the edge is larger than the required resource.

7. An intelligent terminal, characterized in that, intelligent terminal includes: memory, a processor and a hybrid network based service chain mapping program stored on the memory and executable on the processor, the hybrid network based service chain mapping program when executed by the processor implementing the steps of the hybrid network based service chain mapping method according to any of claims 1-6.

8. A computer-readable storage medium, characterized in that the computer-readable storage medium stores a hybrid network based service chain mapping program, which when executed by a processor implements the steps of the hybrid network based service chain mapping method according to any of claims 1-6.

Images